What Is Schema.org? The Shared Vocabulary Behind Structured Data

Schema.org is the collaborative vocabulary for describing entities and their properties in web markup — the shared dictionary that lets a page declare, in terms every major consumer understands, that it describes a SoftwareApplication with an offers price and an aggregateRating. Google, Bing, and Yahoo launched it jointly in June 2011 (Yandex joined months later), a rare act of search-engine cooperation that made one vocabulary the de facto standard instead of three competing ones.

How is the vocabulary organized and governed?

The vocabulary is a type hierarchy rooted at Thing, descending through branches like CreativeWork, Organization, Product, and Event into roughly 800 types carrying over 1,400 properties. Types inherit properties from ancestors — a BlogPosting is an Article is a CreativeWork — so consumers can process specific types they don't fully know by falling back to parents. Since 2015, stewardship has run through the W3C Schema.org Community Group, which develops changes in the open on GitHub and publishes versioned releases at schema.org. The vocabulary is syntax-independent: JSON-LD, Microdata, and RDFa all express it, with JSON-LD now the dominant carrier.

Which subset matters for AI visibility?

AI-era consumption concentrates on entity resolution and answer extraction, which elevates a specific slice of the vocabulary. Organization with sameAs links anchors your brand entity against Wikidata, LinkedIn, and Crunchbase — the disambiguation layer that keeps engines from confusing you with a similarly named company. Article/BlogPosting with author, datePublished, and dateModified feeds credibility and freshness assessment. Product and SoftwareApplication with offers power recommendation-style answers. FAQPage and DefinedTerm mirror the question-answer and definition shapes engines extract verbatim. BreadcrumbList encodes hierarchy cheaply. That handful, implemented accurately, covers most of the machine-facing content work a brand needs.

What does correct usage look like?

Match markup to reality: mark up only entities genuinely on the page, with values identical to visible content, using the most specific applicable type. Connect entities with @id references so your Organization, WebSite, and Article nodes form a small graph rather than disconnected islands. Validate with the schema.org validator and Google's Rich Results Test, and re-validate after template changes — silently broken JSON-LD is the most common decay mode. The vocabulary rewards boring correctness over exotic type usage.